I received my copy of Texas Hold’em Pro 2007 in a no-frills CD sleeve. Installation was a snap and within 3 minutes I was playing with the program. It isn’t the prettiest interface. There is very little in the way of animation and the green and brown color scheme wasn’t much to look at but I was more interested in the guts of the program.
Texas Hold’em Pro is a limit holdem training and simulation software for PCs. The program shows ten available seats positioned around a table. Each player’s AI is completely customizable and you can choose the total number of players you play against. This is nice for someone interested in working on their short-handed or HU game. I decided to grab a seat and start playing without customizing any of the players to see how well this program performed right out of the box. In my opinion, the players play surprisingly well. The software controlled players were check-raising sets, value-betting top pair hands and checking for free cards when appropriate. I didn’t see many flaws in the play.
My next goal was to customize the players to represent a more typical game I might play in. Using the default buttons I made a few players “Tight”, “Loose”, “Aggressive” and “Weak” and again played against my foes. These default models resembled their real-life counterparts very well. I decided to add a new player of my own creation – the maniac. By adjusting a series of intuitive sliders I adjusted the maniac’s playing style. I could precisely control how often he folded on each street and how frequently he bet or raised. After adjusting these slider my newly created maniac did exactly what I asked him to do – play like a hyper-aggressive nit. It was a lot more fun to play with him when there was no real money on the line but I could see this being invaluable practice for anyone who regularly plays against a raising machine.
In terms of play against the computer Texas Hold’em Pro has several different modes. Normal Play mode is just what you would expect. It plays just like you are playing on-line. In Compete Play mode, for each hand you play, a computer player called “Roy” also plays the hand in the background. At the end of each hand, the result of this hand and the total will be shown for both you and “Roy”. This is a good way to consider alternate lines. In Internet Play mode you can choose your hole cards, your opponents hole cards and the actions of your opponents. As you make decisions a window pops up showing you the expectation of each option. In Watch Play mode all hands are played by the computer with all cards revealed to the viewer. You can watch the computer players play one hand at a time for the purpose of learning, or you can let the computer plays continuously for the purpose of simulation analysis. There is one other mode called “stack the deck” where you can continuously deal yourself the same two cards and play against your opponents with randomized distributions.
My favorite poker game currently is short-handed limit poker. I decided to see how well the software could play this game. I limited the field to six player, made a couple players loose and a couple others aggressive and began playing. In this realm I felt the software fell short. There was a lot of checking on the flop rather than the typical continuation bets that are common in this game. There were also a lot more donk bets rather than check-raises which also seemed a little off. Clearly this software is better suited for full-ring players.
Lastly – I decided to test the ability of the software to aide in analysis. I placed a customized player on the button and set him to steal raise with any two cards and then play well after the flop. I set the small blind as a tight player and the big blind as a loose and somewhat aggressive player. I locked the button in place so that every hand would be three-handed and the stealer would always have the button. My goal here was to evaluate stealing hand ranges versus a big blind who defends loosely. Since this is a very common situation in short-handed poker I thought the simulation might shed some light on the hands a button should steal with. I started the simulation and went to bed. In the morning I had more than ten thousand hands worth of data.
The problem with this simulation was making sense of the data. The available data kept track of the statistics for each player which isn’t very meaningful here since the button never moves. The data that the program offers about specific hands is only for the player selected as yourself. Since you cannot customize the playing style of the player selected as yourself I had to set myself as the small blind and therefore this data isn’t meaningful either. Basically there isn’t a good way to run a simulation with meaningful data because the program only tracks hole cards for the player designated as yourself but you cannot automate your play once you set a seat as yourself. So unless you are willing to play 100,000 hands there doesn’t seem to be a good way to run a meaningful simulation.
Overall I think this is a solid program for learning some of the basics of full-ring limit play. It doesn’t translate very well to short-handed play or heads up play and the simulation and analysis side of the software wasn’t well thought out. Hopefully they can improve upon these aspects as I think the full-ring hand play is very good and shows a solid amount of promise. I had one technical problem when installing the software and got a very quick and friendly response and even a follow-up e-mail from their support team. I thought this also deserved some mention as good support is rare these days.